Chapter_10_Field Research Flashcards
(44 cards)
2 Problems of Control in Field Settings
- Control Over Variables
- Control Over Research Populations
Problems with Control Over Variables
- limited control over their independent variable
- recording of observations is especially difficult
- limited control over extraneous variables
Problems with Control Over Research Populations
- natural settings may also be based
on unrepresentative samples - impossible to randomly assign
Field experiments
balance between control and naturalism in research
- by studying people’s natural behavioral responses to manipulated IVs in natural settings
Settings and Sample
Publicness
a wider variety of people in them than are other settings
High publicness
Find any member of the public.
- e.g. public parks, streets, and highways
Middle publicness
People are likely to be similar on one or more characteristics.
-e.g. public meetings, racetracks, areas outside residential houses, and college campuses
Low end publicness
Less public and more institutionalized, where people are linked by some common characteristic
- e.g. student housing, public transportation, restaurants, and stores
Characteristics of a Good Setting
- ability to manipulate the independent variable in the setting
+ random assignment - events could reasonably be expected to take place in that setting
- permission
The Accosting Strategy
select a specific subject who then becomes the target for the experimental intervention
Problems in Field Experimentation
- Construct Validity
- Control Over Extraneous Variables
- Vulnerability to Outside Interference
Construct Validity Problems
- using manifest variables rather than hypothetical constructs
- solution
+ a pilot study - discriminant validity can be a problem
- be influenced by other people
Control Over Extraneous Variables Problems
- IVs are confounded with the environment
- participants characteristics might be confounded with setting
Natural Experiments
- IVs = events outside their control or institutional policies
- correlational study
- no IV manipulatation
- no random assignment
- no control over extraneous variables
Quasi-experiment
- IVs = existing groups
- manipulate IV
- random assignment
- no control over extraneous variables
Nonequivalent Control Group Design
- experiment and control groups are similar
- not randomly assigned
2 Cons of Nonequivalent Control Group Design
- pre-existing differences
- biased selection
The Problem of Pre-existing Differences
Extreme ends
- regression to the mean at post test
The Problem of Biased Selection
Participants’ personal characteristics are confounded with the IVs
- self selection
- after-the-fact selection differences
-> differential attrition
Focal Local Controls
Ensure that the control and treatment groups are as similar as possible
- same location
- share important (focal) personal characteristics
Nested ANOVA
can separate the variance in the dependent variable that is due to
the effect of the independent variable
The Interrupted Time Series Design
- single-case approach
- baseline period
is followed by a treatment that “interrupts” the baseline, - followed by a period of post-treatment observations
Equivalent time samples design
- original design
- withdrawing the treatment
- making observations about behavior without the treatment
- the treatment is reintroduced
- behavior is assessed
Control Series Design
an interrupted time series design with the addition of one or more control cases